ARTICLE pubs.acs.org/Langmuir
Polyethylenimine-Impregnated Mesoporous Silica: Effect of Amine Loading and Surface Alkyl Chains on CO2 Adsorption Aliakbar Heydari-Gorji, Youssef Belmabkhout, and Abdelhamid Sayari* Department of Chemistry, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada
bS Supporting Information ABSTRACT: Poly(ethyleneimine) (PEI) supported on pore-expanded MCM41 whose surface is covered with a layer of long-alkyl chains was found to be a more efficient CO2 adsorbent than PEI supported on the corresponding calcined silica and all PEI-impregnated materials reported in the literature. The layer of surface alkyl chains plays an important role in enhancing the dispersion of PEI, thus decreasing the diffusion resistance. It was also found that at low temperature, adsorbents with relatively low PEI contents are more efficient than their highly loaded counterparts because of the increased adsorption rate. Extensive CO2 adsorption desorption cycling showed that the use of humidified feed and purge gases affords materials with enhanced stability, despite limited loss due to amine evaporation.
’ INTRODUCTION In recent years, extensive research activities have been focused on the development of adsorption-based CO2 capture technologies. The investigated materials include oxides,1 zeolites,2 metal organic frameworks,3 activated carbon,4 and amine-containing materials.5 24 With regard to the latter, two approaches have been developed using most often mesoporous supports. One procedure is the grafting of organosilanes onto largesurface-area materials,5 11 whereas the other is the physical impregnation of the support with amines.12 23 Because of its simplicity, low cost, and the ability to load large amounts of amine, the impregnation technique has been widely used. Although the impregnation of relatively low-molecular-weight amine-containing species such as diethanolamine13 and tetraethylenepentamine17,20,23 led to adsorbents with interesting equilibrium and kinetic properties, they were unstable and showed a steady decrease in the adsorption capacity because of gradual evaporation. Significantly more work dealt with the impregnation of heavier amine-containing species, namely, poly(ethylenimine) (PEI). The importance of the pore size (and pore volume) of the support in the PEI loading and CO 2 adsorption capacity was clearly demonstrated.15 18,23 For example, Son et al.15 reported that at 50 wt % PEI loading, KIT-6 mesoporous silica with 6 nm pores adsorbed 135 mg/g in a stream of pure CO2 at 75 °C versus 111 mg/g when using a 2.8 nm pore size MCM-41 silica with the same amine loading. Hollow mesoporous capsules23 with much larger inner diameters accommodated PEI loadings of up to 83 wt % with an adsorption capacity of as high as 250 mg/g in the presence of pure CO2 at 75 °C. Furthermore, using hexagonal mesoporous silica (HMS) as a support, Chen et al.18 showed that in addition to the intrinsic pore size and volume the pore structure may also play an important role. Nevertheless, r 2011 American Chemical Society
despite the improvement in adsorption capacity via the proper selection of a support, most of these materials were dominated by diffusion resistance and showed an optimum adsorption temperature at about 75 °C,15 19,23 but little attention was paid to adsorption at around room temperature. To address the negative effect of diffusion resistance, some workers decreased the viscosity of supported PEI by the addition of polyols.21,22 Here we describe an innovative approach to enhancing the effectiveness of the adsorbent via an improvement of the PEI dispersion using as a support a large-pore silica with a surface layer of long hydrophobic alkyl chains. We also demonstrated that for adsorption at low temperature, materials with low PEI contents perform much better than highly loaded adsorbents.
’ EXPERIMENTAL SECTION Synthesis of Materials. All chemicals were obtained from SigmaAldrich and used as supplied. A detailed description of the synthesis of pore-expanded MCM-41 silica (PE-MCM-41) and its structural characteristics has been reported elsewhere.25 27 Briefly, MCM-41-type silica was synthesized at 80 °C using cetyltrimethylammonium bromide (CTAB) as a structure-directing agent and a 25% solution of tetramethylammonium hydroxide (TMAOH) in water for pH adjustment. Pore expansion was carried out through the hydrothermal treatment of as-synthesized MCM-41 using N,N-dimethyldecylamine (DMDA) as a swelling agent at 120 °C for 72 h. As-synthesized PE-MCM-41 was either calcined at 550 °C to remove both the surfactant template and pore-expanding agent or extracted with ethanol to remove only the Received: May 12, 2011 Revised: September 8, 2011 Published: September 08, 2011 12411
dx.doi.org/10.1021/la202972t | Langmuir 2011, 27, 12411–12416
Langmuir
ARTICLE
Figure 1. Schematic representation of the syntheses of PMC and PME. swelling agent, giving rise to PMC and PME, respectively. As for the preparation of CO2 adsorbents, the required amount of PEI (Aldrich, average Mn ≈ 423) was dissolved in methanol before being added to the mesoporous silica support. The resulting slurry was continuously stirred at room temperature until the solvent was evaporated, and then the sample was dried at 60 °C under reduced pressure (700 mmHg). The obtained adsorbents were designated as PMC-PEI(x) and PME-PEI(x), where x represents the weight percent of PEI in the adsorbent. Material Characterization. All materials were characterized by nitrogen adsorption desorption at 196 °C using a Micromeritics ASAP 2020 automated volumetric instrument. The specific surface area (SBET) was determined using the BET method in the 0.05 0.2 relative pressure range, and the pore size distribution (PSD) was calculated using the KJS (Kruk Jaroniec Sayari) approach.28 The pore diameter (DP) corresponds to the maximum of the PSD, and the total pore volume was calculated from the amount adsorbed at a relative pressure of about 0.99. The 13C CP/MAS NMR experiments were conducted on a Bruker Avance 500. The spinning frequency was set to 10 kHz. The contact time was 2 ms, with recycle delays of 2 s. The time for data collection ranged from 7 to 16 h. The actual amount of organic compounds on the materials was measured before and after ethanol extraction via thermal decomposition using a thermogravimetric analyzer (Q500, TA Instruments). The sample was dried at 100 °C and heated to 800 °C in flowing nitrogen at 1 °C/min and then to 1000 °C in air. The weight losses were used to calculate the relative amounts of DMDA and cetyltrimethylammonium (CTMA). Adsorption Measurements. Adsorption tests were carried out using a thermogravimetric analyzer (TGA) connected to a gas-delivery manifold. A sample was treated in flowing N2 at 100 °C, cooled to the desired adsorption temperature, and then exposed to flowing CO2 (90 mL/min) for 180 min. The stability of a selected sample, PMEPEI(50), was evaluated using cyclic adsorption desorption measurements in the presence of dry and humidified feed and purge gases. The sample was first pretreated in flowing dry N2 at 100 °C and then cooled to 75 °C before switching to pure CO2 for 30 min. Desorption took place at the same temperature under pure N2 for 30 min. In the case of humid conditions, the feed and purge gases used for adsorption and desorption were bubbled through a temperature-controlled water saturator. The 6% RH used in this work corresponding to the vapor pressure at 20 °C was selected for convenience. In this case, the adsorption and regeneration times were 20 and 45 min, respectively. A total of 120 adsorption desorption cycles were achieved in both cases.
Figure 2. 13C CP/MAS NMR spectra of as-synthesized PE-MCM-41 and PME.
’ RESULTS AND DISCUSSION Figure 1 shows a schematic representation of the synthesis of ethanol-extracted and calcined pore-expanded MCM-41. As shown in Figure 1, as-synthesized PE-MCM-41 exhibits large channels (of up to 20 nm) containing a layer of cetyltrimethylammonium cations (CTMA), with the ammonium groups interacting with the pore wall surfaces, as well as swelling agent dimethyldecylamine, (DMDA).25,26 As demonstrated earlier,27 the selective extraction of DMDA by organic solvents such as ethanol affords a highly porous material with a layer of CTMA cations on the channel surfaces, resulting in highly hydrophobic silica. Both calcined and ethanol-extracted PE-MCM-41, referred to as PMC and PME (Figure 1), were used as supports for PEI. Figure 2 shows the 13C CP/MAS NMR spectra for as-synthesized PE-MCM-41 and PME. Two peaks at 45.7 and 60 ppm in the spectrum of as-synthesized PE-MCM-41 were assigned to the methyl groups attached to N and the first carbon atom (2*) of the alkyl group in DMDA, respectively. 29 These peaks disappeared after ethanol extraction in the spectrum of PME, 12412
dx.doi.org/10.1021/la202972t |Langmuir 2011, 27, 12411–12416
Langmuir
ARTICLE
Figure 4. CO2 uptake versus temperature for (a) PME-PEI(55) after 180 min, (b) PME-PEI(55) after 30 min, (c) PMC-PEI(55) after 180 min, and (d) PMC-PEI(55) after 30 min of exposure to pure CO2.
Figure 3. Decomposition profiles for (a) as-synthesized PE-MCM-41 and (b) PME.
Table 1. Structural Properties of Materials materialsa
SBET (m2/g)
Vp (cm3/g)
PMC
1254
2.44
PME PME-PEI(20)
570 269
1.59 0.68
PME-PEI(30)
16.7
0.04
PMC-PEI(55)
37.4
0.09
a
PME with more than 50 wt % PEI did not show any surface area or pore volume.
indicating the complete removal of DMDA and the retention of CTMA species. The thermal decomposition profiles of as-synthesized PEMCM-41and PME are shown in Figure 3. The rate of weight loss for as-synthesized PE-MCM-41 (Figure 3a) exhibited two maxima at 115 and 220 °C, corresponding to the release of DMDA and CTMA, respectively. In contrast, the thermal decomposition profiles of PME (Figure 3b) exhibits a single feature with a
maximum weight loss rate at 223 °C, indicating the occurrence of CTMA only (i.e., no DMDA was left after ethanol extraction). The nitrogen adsorption desorption isotherms for PMC, PME, PME-PEI(20), PME-PEI(30), and PMC-PEI(55) materials are shown in Figure S1 (Supporting Information). The pore sizes of PMC and PME were 11.4 and 10.9 nm, respectively. Table 1 shows the surface area and pore volume of selected materials. The diminishing surface area and pore volume of the materials upon PEI impregnation provide strong evidence that PEI is located within the channels. Notice that PME and PMC, like most periodic mesoporous silicas,30 have very small external surface areas compared to the internal surfaces associated with their pore systems. Therefore, if any PEI does not migrate inside the channels, the material will readily turn into a pastelike substance. No such behavior has been observed. Figure 4 shows the CO2 uptake for PMC-PEI(55) and PMEPEI(55) in the presence of 100% CO2 at different temperatures for 30 and 180 min. Although PME had a lower pore size and pore volume than PMC (Table 1), PME-PEI(55) exhibited much higher CO2 uptake than did PMC-PEI(55) at all temperatures. For instance, at the optimum temperature (i.e., 75 °C) the CO2 uptake for the PME-based adsorbent was 2.3 times higher than for its PMC counterpart. Moreover, at 75 °C and beyond, CO2 adsorption on PME-PEI(55) reached equilibrium within 30 min of exposure because the CO2 uptake did not increase upon further exposure up to 180 min. In contrast, in the presence of PMC-PEI(55), the equilibrium adsorption capacity was not achieved even at 100 °C. This behavior indicates that the CO2 adsorption is controlled mostly by diffusion, which reflects the state of dispersion of PEI within the pore system. On the basis of the experimental findings, it is inferred that at the same loading, PEI exhibits a lower diffusion resistance on PME than on PMC and thus a higher dispersion. For illustration, Table S1 (Supporting Information) provides direct evidence that the CO2 diffusivity in PME-PEI(55) is greater than that in PMCPEI(55). The different behaviors of PMC-PEI(55) and PME-PEI(55) may be associated with the nature of their pore wall surfaces. Indeed, PMC, whose surface is populated by hydroxyl groups (Figure 1), is rather hydrophilic, whereas the surface of PME is laden with a layer of long hydrophobic hydrocarbon chains (CTMA). 12413
dx.doi.org/10.1021/la202972t |Langmuir 2011, 27, 12411–12416
Langmuir
ARTICLE
Figure 6. Comparison of dynamic pure CO2 uptake after 180 min of exposure at 25 °C for (a) PME-PEI(30) and (b) PME-PEI(50). Figure 5. Literature survey of the CO2 adsorption performance of PEIcontaining mesoporous silica adsorbents at 75 °C in the presence of pure CO2. (The highest uptake reported in each reference has been considered here.)
Nonetheless, although PME does not disperse in water even after extensive stirring, it readily sinks in methanol. It is inferred that during the synthesis of PME-PEI(x) adsorbents the PEI methanol solution migrates inside the pores and upon evaporation of the solvent the PEI penetrates the alkyl chain network and disperses therein. It will thus exhibit higher dispersion with diminished diffusion resistance, leading to faster and higher CO2 uptake. It is conceivable that the layer of long alkyl chains form a porous network within the silica channels, consistent with enhanced PEI dispersion. However, the question of whether the dramatic improvement in CO2 uptake (Figure 4) and adsorption kinetics (Figure S4) is related to the surface hydrophobicity per se or to the network of surface porosity made with long alkyl chains will be delineated in future work. On the contrary, on the bare silica surface of PMC, PEI is likely to agglomerate inside the channels, giving rise to a higher diffusion resistance thus much slower adsorption kinetics. The optimum temperature for CO2 adsorption of PME-PEI(55) was 75 °C because actual equilibrium was achieved, whereas the CO2 uptake on PMC-PEI(55) increased beyond 75 °C (i.e., it did not reach thermodynamic equilibrium). A comparison of CO2 uptake in the presence of the current PME-PEI(x) adsorbents under selected conditions (75 °C and 100% CO2) with literature data provides further evidence of the importance of the nature of the support surface. As shown in Figure 5, PEI-supported mesoporous silicas such as MCM-41,12 KIT-6,15 SBA-15,16 silica monolith,17 HMS,18 and mesoporous silica capsules23 exhibited a lower CO2 adsorption efficiency (amount of CO2 adsorbed per gram of PEI) than did PMEPEI(x) at any loading, even for supports with larger surface areas and pore volumes. Thus, the support pore structure is not the only factor affecting the CO2 adsorption capacity because the PEI dispersion inside the support channels can significantly enhance the adsorbent performance. To achieve high CO2 uptake, workers most often used highly loaded materials.24,31 However, as the PEI content increases, the diffusion resistance also increases, adversely affecting the CO2 uptake, particularly at low temperature. For example, although
Figure 7. CO2 uptake vs temperature for (a) PME-PEI(55), (b) PMEPEI(50), (c) PME-PEI(30), and (d) PME-PEI(20) after 180 min of exposure to 10:90 CO2/N2.
highly loaded PME-PEI(x) reached its equilibrium capacity quite readily at ca. 75 °C, it was also dominated by slow diffusion at low temperature (Figure 4, traces a and b). Figure 6 (trace b) shows that in the presence of pure CO2 at 25 °C the uptake for PMEPEI(50) after 180 min of exposure did not exceed 6.5 versus 18% under the same conditions at 75 °C. Nonetheless, at such a high loading, PEI supported on conventional mesoporous silicas hardly adsorbs any CO2 at room temperature.15,19 To circumvent the adverse effect of the diffusion limitation at low temperature, it is recommended to use adsorbents with lower loadings. Under such conditions, the CO2 uptake will be controlled by the much faster adsorption kinetics associated with the increased dispersion of PEI at low content, thus diminishing the diffusion resistance. An example is shown in Figure 6. As seen, under the selected conditions, PME-PEI(30) showed significantly higher and faster CO2 uptake than did PME-PEI(50). Other supported PEI adsorbents may exhibit similar behaviors, regardless of the nature of the support. Figure 7 shows the CO2 uptake for the PME-PEI(x) adsorbent with x = 20, 30, 50, and 55 after 180 min of exposure to a 10:90 CO2/N2 mixture at different temperatures. As seen, for 12414
dx.doi.org/10.1021/la202972t |Langmuir 2011, 27, 12411–12416
Langmuir
ARTICLE
and pressures. It was demonstrated that in addition to the pore size, volume, and structure, the actual nature of the silica surface plays a key role in the performance of supported PEI for CO2 adsorption. PEI seems to disperse better in an alkyl surface layer, leading to enhanced CO2 uptake. As an example, it was found that PEI impregnation on solvent-extracted, pore-expanded MCM-41 mesoporous silica whose surface is covered with a layer of long-chain alkyltrimethylammonium cations affords much more efficient adsorbents than does the corresponding calcined silica. Nonetheless, at high loading, CO2 adsorption at low temperature (
